J C Mateus-Solarte1, R Carvajal-Barona. 1. División Salud, Fundación para la Educación Superior Social, Cali, Colombia; Universidad del Valle, Cali, Colombia. jcmateus@fundacionfes.org
Abstract
BACKGROUND: Early diagnosis and treatment are fundamental to tuberculosis (TB) control. Nevertheless, the effectiveness of TB management continues to be influenced by treatment adherence. OBJECTIVE: To determine which factors are predictive of adherence to TB treatment at the time of diagnosis in Colombia. METHODS: A cohort of 300 patients newly diagnosed with TB was followed up over 26 weeks. Treatment adherence was measured by determining whether the patient took all or part of the 84 doses in the 26 weeks of treatment. A logistic analysis was carried out and the predictive power of the final variables was determined by means of a receiving operator curve analysis. RESULT: A high incidence of partial completion of treatment (65.6%) was found. Significant associated factors were 1) living away from the family, 2) overcrowding at home (> or =2 persons per bedroom), 3) lack of family support, 4) living >10 min away from the treatment facility and 5) not having used the services of the treatment facility before. CONCLUSION: Several factors can be measured on PTB diagnosis that would help identify those patients at higher risk for treatment non-adherence. The predictive value of each of these factors alone was weak, but if associated their predictive value was high.
BACKGROUND: Early diagnosis and treatment are fundamental to tuberculosis (TB) control. Nevertheless, the effectiveness of TB management continues to be influenced by treatment adherence. OBJECTIVE: To determine which factors are predictive of adherence to TB treatment at the time of diagnosis in Colombia. METHODS: A cohort of 300 patients newly diagnosed with TB was followed up over 26 weeks. Treatment adherence was measured by determining whether the patient took all or part of the 84 doses in the 26 weeks of treatment. A logistic analysis was carried out and the predictive power of the final variables was determined by means of a receiving operator curve analysis. RESULT: A high incidence of partial completion of treatment (65.6%) was found. Significant associated factors were 1) living away from the family, 2) overcrowding at home (> or =2 persons per bedroom), 3) lack of family support, 4) living >10 min away from the treatment facility and 5) not having used the services of the treatment facility before. CONCLUSION: Several factors can be measured on PTB diagnosis that would help identify those patients at higher risk for treatment non-adherence. The predictive value of each of these factors alone was weak, but if associated their predictive value was high.